CSE571: Reinforcement learning project submission

["Tuan A. Nguyen" <tanguye1@asu.edu>]

Hi all:

The implementation for reinforcement learning will be due on this Wednesday. Please send me your source code and reports by the end of the date, and help me title your email starting with CSE571-RL.

The report should clearly explain your observations (preferably with graphs) and how you explain the results. Even in the case that you don't see convergence, for example, does it help to increase the minimum number of times each actions should be tried, Ne? What's about R+? Do you have any modifications on the exploration functions to help Q-learning converge faster?

You may also want to describe the data structures you used and the pseudo-code that you implement (I got some emails asking steps of the algorithms, so just want to make sure you're doing Q-learning).

Thanks,

Tuan